Below are the 20 most recent journal entries recorded in wtogami's LiveJournal:

[ << Previous 20 ]

Sunday, June 28th, 2009
9:42 pm	pidgin-2.5.8 several important bug fixes http://kojipkgs.fedoraproject.org/packages/pidgin/2.5.8/ pidgin-2.5.8 builds for Fedora 9, 10, 11 and 12.  Please let me know if anything became worse. (Comment on this)
Saturday, June 20th, 2009
11:55 pm	pidgin-2.5.7 fixes Yahoo protocol pidgin-2.5.7 was released fixing the sudden Yahoo protocol change.  Builds for Fedora 9, 10, 11 and rawhide are available for testing here.  Please let me know if you see any regressions.  RHEL-4 and RHEL-5 builds are here. (2 Comments |Comment on this)
Thursday, June 11th, 2009
8:46 pm	Tivo HD upgrade to 1TB Late last year I attempted to upgrade my Tivo HD hard drive.  It came with a tiny 160GB disk (only 20 hours HD video).  I first tried the Seagate 7200.3 DB35 750GB which turned out to be a dismal failure.  This drive according to Seagate's documentation was supposed to be designed for DVR's.  Unfortunately, it seems to be a terrible drive.  The first drive I received burnt out instantly as a IC on the bottom fried itself at first power-on.  The second replacement worked, but turned out to be entirely unstable for use in the Tivo.  The video would frequently freeze, and I would hear loud clicks from the drive as it took several seconds to recover.  When it did work it inserted lots of data corruption in the video stream.  Sometimes the errors caused the Tivo's tuner to give up.  Sometimes the entire Tivo would just crash.  To make matters worse, the drive would become very hot to the touch.  Seagate's own diagnostic utility says nothing is wrong with the raw drive, and the drive was not returnable. Last week I bought a Western Digital Caviar Green 1TB WD10EADS.  This guy pointed at similar WD drives better designed for DVR's, but the EADS was the cheapest and most likely to be useful outside a DVR if it turned out to not work for the Tivo. This 1TB drive seems to be working great in the Tivo HD.  No apparent problems handling recording two HD streams while watching a previous HD recording.  The device reports "142 hours" of HD video.  Also to my surprise it is both quieter and cooler to the touch than the original Tivo HD WD SATA2 disk.  It is only a week with this drive, but so far so good. UPDATE: Adam D. Ligas pointed out that if you use non-DVR drives like the WD10EADS you should probably adjust its acoustic settings.  While the WD10EADS was already quieter and less power hungry than the original Tivo drive, apparently it can be even better.  He also pointed out the "Supersize" option which can squeeze 157 hours of HD instead of 142 out of a 1TB drive.  You need Windows XP SP2+ and Winmfs to do it.  Apparently the Tivo ships with 10% of the drive reserved for Tivo advirtisments.  There reportedly is no drawback to turning on Supersize mode where most of that space is allowed to be used for normal recordings.  Also this giant FAQ seems to indicate that the WD10EADS will not work with the earlier Tivo Series 3, although the similar DVR-specific drives like WD10EVCS or WD10EVVS do work.  It does however work with the newer Tivo HD, so I guess I lucked out. Adam wrote the below to set the acoustic settings on a drive: #1 - Use hdparm to adjust the acoustic settings on the drive. hdparm -M /dev/sdX (where X is the TiVo drive assignment) If the value comes back as anything other then 128 (the lowest) do: hdparm -M 128 /dev/sdX (where X is the TiVo drive assignment, again) Your TiVo drive will be even more quiet then it already is. (Comment on this)
Tuesday, April 28th, 2009
10:50 pm	DVD installer on USB stick http://wtogami.fedorapeople.org/temp/livecd-iso-to-disk.sh I modified the livecd-iso-to-disk.sh script to automatically detect if you specified the DVD iso installer.  If so, it copies it to the target USB stick and configures syslinux to use it as the install source for Anaconda.  Note: This will work with only Fedora 11 Preview DVD iso and higher.  Patch was submitted for review to the livecd-tools package.  Hopefully some version of this functionality will make a future version. Please let me know if you run into any problems with this script. UPDATE: Jeremy decided this does not belong in the official livecd-iso-to-disk.sh script.  That's OK.   This can eventually become its own script and functionality added to liveusb-creator.  For now this script should work for Fedora 11 DVD ISO's. (3 Comments |Comment on this)
Sunday, April 26th, 2009
12:02 am	Chrono Cross woes on PS3, copying PSOne saved game from PS3 Last year I bought the classic Playstation One game Chrono Cross.  It is one of the greatest games of all time, like its predecessor Chrono Trigger for the Super Nintendo.  Gamespot rated Chrono Cross a perfect 10 out of 10. To my horror, I found that the game is effectively unplayable due to bugs in Playstation 3's emulation.  The Net has scattered reports like this of anguished cries ignored by Sony.  Apparently Sony has long ago stopped development of the PSOne and PS2 emulation on the PS3, so this bug will likely never be fixed.  The bug seems to be a race condition that triggers when your characters or an enemy casts a spell.  Some battlegrounds are heavily animated, causing the PS3's PSOne emulation to be bogged down.  You see the frames slow and the music plays slowly.  If the spell casting animation happens during this slow-down, the game is highly likely to get stuck.  I tried to play through the game anyway.  Some parts of the game can be brute forced by playing it 50 times and by luck you might be able to get through a battle without the race condition deadlocking the game.  But some parts like the S.S. Invincible or obtaining the Einlanzer are nearly impossible because of extended length battles and heavy background animation.  Frustrated with the broken Playstation 3 emulation, I gave up entirely for a few months. I eventually figured out that it was possible to run the game that I bought and paid for on the pSX emulator.  I suppose other emulators like ePSXe might work better, but ePSXe for Linux hasn't been updated since 2003 and is thus ancient and broken.  pSX is not without problems, it gets confused by modern ALSA and pulseaudio, but it can be made to work with some workarounds.  It is sad that none of these PSX emulators are open source.  It seem this particular bug could be easily fixed if source were available.  I am excited to discover that the PS3 controller plugged in via USB is automatically is now automatically detected by Linux and configured as a joystick.  You can easily map the buttons of the PS3 controller to the virtual Dualshock in pSX and things thereafter just work.  It seems somebody wrote drivers to use the PS3 controller in Bluetooth wireless mode with Linux as the host.  Check out the videos on that page.  Pretty cool. http://togami.com/~warren/guides/ps3-psone-saved-game-to-pc-howto/ Now that Chrono Cross was seemingly working in the  pSX emulator, I wanted to copy my saved game file from the PS3 to the PC.  I wrote the procedure to do that on this page.  This page really exists only to praise the timeless awesomeness of Chrono Cross. (2 Comments |Comment on this)
Monday, February 9th, 2009
5:18 pm	Kabukibear83 Youtube account Deleted http://www.youtube.com/Kabukibear83 This guy (Justin Lincoln?) had over 50 great original recordings of himself playing guitar of various songs that he arranged over the past few years.  His entire account is now gone without any explanation.  I am filled with deep sadness.  Does anyone know how to contact him?  He inspired me to buy a guitar and take lessons with the hope to be able to play and arrange songs like him one day. UPDATE: It seems many other people are shocked and saddened by Kabukibear83's disappearance. (2 Comments |Comment on this)
Tuesday, December 16th, 2008
2:54 pm	Netbooting old PPC Macs as Thin Clients UPDATE: I eventually figured out a complicated dhcpd.conf recipe that should theoretically netboot any Apple PPC Mac.  This was released in ltsp-5.1.44.  See k12linux.org. Schools have thousands of old G3 iMacs that are increasingly too slow to be useful. They would be useful again if they were LTSP thin clients, logging into a modern x86_64 terminal server. So I tried to make them boot as thin clients using LTSP and standard Fedora ppc components. The next release of ltsp will automatically build a thin client chroot in /opt/ltsp/ppc with the ltsp-build-client command.  The idea is that you could copy that /opt/ltsp/ppc chroot to your x86_64 server that handles ordinary i386 client netboot, and through dhcpd.conf trickery it could serve netboot to the PPC clients.  Unfortunately I couldn't get it to work. http://en.wikipedia.org/wiki/IMac_G3 The hardware I tried to get netbooting was the slot loading dark blue iMac G3. yaboot-1.3.14-8.fc10 kernel-2.6.27.7-134.fc10 OFW seems incapable of downloading yaboot from a tftp directory. I need to put it directly in the tftp root for it to download successfully. Booting Fedora 10's kernel and initrd, it gets stuck with: runaway loop modprobe binfmt-4c46. A few Google hits, cebbert and dwmw2 agree that this means it is trying to load elf.ko... but elf.ko is built into the kernel so it shouldn't be an issue. dwmw2 thinks the download is being truncated, although this is uncomfirmed. I tried shriking the 7MB initrd down to 3.5MB by removing all unnecessary drivers, but it still fails with this error message. I tried wrapping the vmlinuz and initrd into a zImage, and OFW failed while downloading it with "LOAD_SIZE is too small". zImage seems to lack a way to acquire kernelopts over the network so this wouldn't be a usable method anyway. dwmw2 said to try the F-9 version of yaboot because there were tftp related changes in F-10. No difference. boot enet:0 always fails to attempt to download yaboot from the specified next-server in dhcpd.conf. Only boot enet:<IPADDRESS>,yaboot will download and run yaboot.  Here is the dhcpd.conf that I am using. #1, #3, and #5 seem to be firmware problems with the ancient iMac, but it failed to boot in different ways when I tried to netboot a fairly recent G5 tower.  This points to problems in our software instead of only firmware issues. I am giving up for now. If anybody have better suggestions I might try them later. (1 Comment |Comment on this)
Tuesday, December 2nd, 2008
11:56 pm	Lenovo T60 upgrade to SSD? I wish to upgrade my Lenovo T60 to SSD.  http://www.newegg.com/Product/ProductReview.aspx?Item=N82E16820167005 This review indicates that someone with a Lenovo T60 failed to get Intel's 64GB MLC SSD to work in his Thinkpad.  This makes me wonder, does Lenovo have restrictions that disalllow the use of "non-approved" by Lenovo devices?  We already have to use the cmos hack to use 3rd party wireless cards in a Thinkpad, for example. Anyone know if Lenovo truly does lock down the hard drive interface to only certain "approved" drives? (1 Comment |Comment on this)
Sunday, November 9th, 2008
11:26 am	Full Report: LTSP Hackfest Maine 2008 The Linux Terminal Server Project (LTSP) enables Linux distributions to boot entire networks of diskless clients, and serve Linux desktop as a thin client, with storage handling (USB, CD, floppy) and remote sound.  LTSP is especially attractive for businesses and schools because it is the most cost effective way to deploy large numbers of desktops with extreme reliability and security, with minimal staff necessary to maintain it. This is the report from LTSP Hackfest Maine 2008.  Last year we held this event at the same location, at the cozy Seawall Motel in Southwest Harbor, Maine.  This year's event and grand lobster feast was sponsored by DisklessWorkstations.com, the founders of the LTSP project nearly a decade ago.  Attendees this year include Red Hat, Disklessworkstations.com, LTSP upstream, Resara, Revolution Linux, Symbio, and a few educators and independent consultants that support LTSP at businesses and schools. LDM: LTSP Display Manager LDM is a graphical front-end greeter login screen for ssh.  Thin clients can boot and run LDM, where it asks for the login name and password.  Optionally the user can choose Session type (GNOME, KDE, XFCE, Sugar, etc.) or login language.  Gideon Romm, Scott Balneaves, Stéphane Graber and Francis Giraldeau focused their efforts on LDM improvements where they made much progress.  They implemented proper event handling by splitting X launching out of the LDM binary.  LDM is now a client of xinit where signal handling is now handled.  This eliminated a long standing race condition where LDM would crash, or keyboard layout setting would fail, because they attempted to run before X had fully initialized.  They fixed Tab key handling, where LDM now cycles through username and password dialogs in the same manner as GDM.  Proper syslog usage was implemented, which should make it more consistent to log and debug LDM in the future.  The long standing "kill -1" hack was finally eliminated after more than a year of effort.  LDM before had to forcibly kill the ssh session at logout.  By cleanly allowing the connection to exit, it is now possible to run other post-logout scripts to cleanup the session properly. Further improvements to LDM were discussed and planned: Currently LDM has a significant problem in the way it talks to the remote SSH server.  It currently uses an expect-like parsing chat to determine if it had successfully logged in, or if password expiry and forced password changing is needed.  This is highly problematic because the chat from the SSH server can be localized in any arbitrary language.  The LTSP developers are thinking about switching LDM from parsing chat to a library interface like libssh2.  It is a shame that openssh does not provide a library interface like this.  libssh2 might work, except there are some concerns about its maintainbility given that the upstream project does not appear to be very active and it would be yet another crypto interface with potential security implications.  Debian does not ship libssh2 in the core repository.  Fedora apparently ships libssh2 in core because it is a dependency of libcurl, but this is suspected to be an accident.  RHEL5 does not have libssh2 at all.  More investigation is required into this issue. Handle password expiry better. Handle cases where client does not already know a SSH server key. Chris and Brendan of Resara are working on a qt greeter for LDM.  They also designed an improved protocol for greeter to LDM backend communication. RHEL5 Backport Warren Togami and Bryan Smith of Red Hat focused mostly on creating a plan for backporting LTSP5 to RHEL5.  The following is the basic implementation plan. mkinitrd (Warren): After examining a few different approaches including libkudzu, LTSP's old pci detection code, or system-config-netboot, we settled upon a backport of Fedora's mkinitrd loadDrivers function to RHEL5.  RHEL5's modprobe seems capable of handling modalias based loading.  It appears possible to pull in modprobe and minimal libraries needed to use modprobe without effecting anything else in RHEL5 mkinitrd.  The only difference that we will need to implement is a way to detect which kernel modules are network drivers.  Fedora's mkinitrd can use --with-avail==networking, but this is not possible on RHEL5 because there is no modules.networking group.  We will likely implement something that looks at PCIID's within kernel modules and automatically pulls in matching kernel modules. livecd-tools (Warren): Warren will backport necessary changes from Fedora's livecd-tools to the EPEL5 version to handle client chroot creation. GNOME backports (Warren): Warren will backport a few simple patches in modern GNOME and a few patches from Ubuntu to correct a few annoying problems with GNOME on a multi-user terminal server. Problem: RHEL5 lacks fuse kernel module (Bryan): A full backport to RHEL5 will require the fuse kernel module on both the client chroot (sshfs) and server (ltspfs).  Bryan will work on the business case to achieve this, while Warren will talk with the filesystem engineers about technical feasibility of shipping it. fuse setuid or fuse group permission issue (Bryan): The fuse userspace tools needs to either have setuid or users must be in a fuse group in order for fuse to work on RHEL5.  Fedora 9+ no longer has this problem because it uses gvfs. alsa-plugins-pulseaudio missing from EPEL5 (Bryan): Bryan plans on figuring out the ideal version of alsa-plugins to put a tested version of the alsa-plugins-pulseaudio in EPEL5. Miscellaneous Stuff ltsp-cluster is an optional add-on project from Revolution Linux.  ltsp-cluster enables the ability to have a cluster of LTSP servers that load balance and fail-over.  There is also central configuration management of thin clients, PXE configuration management.  It is a work in progress to get all of this merged upstream. Local apps menu integration merged into ltsp-trunk.  It automatically generates an XDG compliant menu entry for local applications to show up automatically on the menus of the remote GNOME/KDE desktop.  This requires more polish. Serial printer support was merged into jetpipe.  Testing is encouraged. Local Volume Manager is planned.  This would modify ltspfs to support local apps to be able to directly access local devices (like USB sticks). Gideon Romm plans to push his timezone and time server handling code, and PTP Camera support for local devices upstream.  DPMS power saving support should also go upstream soon. Chuck Liebow and David Johnston worked on documentation. Warren began thinking about PXE to Flash update architecture. Fedora K12Linux Live Server was demonstrated. Gideon Romm demonstrated his Symbiont Boot Appliance. Warren fixed a few Fedora 10 LTSP issues. Next LTSP Event The next LTSP Event is planned in Brazil during 2009 at the 10th anniversary of FISL, which is also the 10th anniversay of the LTSP project.  Then the group plans on meeting again at Southwest Harbor for another Maine hackfest during the same time early November 2009. (2 Comments |Comment on this)
Thursday, October 16th, 2008
5:45 pm	Flash 10 for x86 Linux Workarounds Yesterday Adobe released Flash 10 for x86 Linux. It can be stable on Fedora if you follow a number of workarounds and update a few packages. I have documented the latest workarounds and tips at this site: http://macromedia.mplug.org/ http://wtogami.livejournal.com/28380.html As mentioned earlier, further bug fixes ensue at the new nspluginwrapper-devel-list to make Flash 10 and other browser plugins more stable. Please report any problems you see to this list. (2 Comments |Comment on this)
Sunday, September 14th, 2008
10:30 pm	nspluginwrapper Devel List https://www.redhat.com/mailman/listinfo/nspluginwrapper-devel-list NSPluginwrapper Development discussion with the goal of isolating issues and collaboratively working on solutions should go on this list. There was some interest from other Linux distributions and even Adobe to cooperate on the future of nspluginwrapper development. Today it works fairly well, but there remain a number of issues where it is unclear if it is firefox, nspluginwrapper or a particular plugin's fault. The last few times I wanted to blame Flash for a problem it turned out to be firefox, nspluginwrapper or Fedora's NSS port of curl's fault. Isolating the problems to the true source is really important. Related Issues: Bug 456017: Adobe Reader pretty broken in nspluginwrapper nspluginwrapper Windowless plugin: After using all the workarounds documented here, Adobe Flash Player 10 is still broken on some sites in nsplugniwrapper, while it works unwrapped. For example, Youtube videos embedded on Google News fails after a few seconds of playing in nspluginwrapper unless you use the WindowlessDisable=true workaround. Other nspluginwrapper issues: Other websites fail even with the WindowlessDisable=true workaround. I need to remember to write down addresses when I hit these again... Gwenole hopes Adobe can suggest a better way to proxy the OpenGL directives than the ugly hacks nspluginwrapper-1.1.0 currently uses. Gwenole, could you please describe exactly what you need on the list after everyone has joined? The list could use 100% reproducer procedures with far less vagueness than these examples. Those test cases are needed to narrow down the true cause of various problems. (Comment on this)
Saturday, August 23rd, 2008
10:21 pm	Wanted: 20k+ Tiny Linux Tiny Diskless Workstations I am currently searching for the ideal thin client/diskless form-factor mini computer with Intel Atom. This research is to make hardware recommendations for a school system who may deploy tens of thousands of Linux diskless workstations. Dear Manufacturers: Hardware Samples Needed During the past year, various vendors have been sending me thin client hardware for development and testing of Fedora LTSP5. The hardware samples have been very useful in fixing various issues in the kernel and X to better support the hardware. While I do not make any guarantees, generally all hardware sent to me eventually became usable with Fedora LTSP5. I currently cannot recommend any particular brand or model of Intel Atom-based tiny computer for mass deployments of Linux diskless workstations. I don't have the budget to purchase different brands of thin clients to test suitability for mass deployments as Linux diskless workstations. I acquired an Asus EEEPC 901 in order to do general Fedora portable and Fedora LTSP5 testing on the Intel Atom platform. But testing one particular machine is not good enough. I've found the BIOS of various Atom machines to have bugs, so specific brands and model numbers must be tested for compatibility and working PXE netboot. To manufacturers and vendors: If you want your specific machine to be part of my test cycle for current and future versions of Fedora and RHEL, please send me hardware. Read here for my standard disclaimer regarding hardware samples. Contact me at wtogami at redhat dot com. Below are my notes regarding Intel Atom based machines for diskless client use and how things I see on the market compare to the requirements. Some might ask, why am I so enthusiastic about Intel Atom hardware specifically? At the bottom is some background information of why. Ideal Specifications for Tiny Linux Diskless Workstation 2GB RAM Video with excellent Linux 3D support (currently only Intel) Small and sleek size, we do not need disks Dual-core Atom when it becomes available Nice to have: Fanless (noiseless) operation Nice to have: SD/MMC/MS/etc front-loading slot Nice to have: VGA or DVI option TranquilPC T7 Thin Client This is one of the first Intel Atom-based thin clients that I've seen on the market. The specifications sound very good although I have not had the opportunity to test this machine. I wish I could find an American vendor of something similar to this. Pros: No fan! Absolutely zero noise. No moving parts reduces likelihood of hardware failure. (Likely the entire case acts as a heatsink.) Cons: No SD slot like the competitors, seems to have only VGA. Asus Eee Box B202 The desktop version of Asus' successful Eee PC netbook. I also hear that a version of this embedded into a 19" or 21" LCD monitor is planned. Both would seem to be very attractive for mass deployments of Linux diskless workstations with only *minor* alterations for bulk purchase, like 2GB instead of 1GB RAM. Pros: For slightly more than the TranquilPC T7 you get an even sleeker case, SD slot, hard drive and wireless card. DVI is a plus. Cons: Is it possible to buy in bulk quantities without a hard drive or wireless card? Hard drive generates extra heat and noise, and it is a moving part which increases the potential for hardware failure. This has a cooling fan (unlike the TranquilPC). Diskless workstations will want 2GB RAM instead of 1GB. The price is still attractive with this standard listed model, but buying it in bulk quantities with 2GB RAM would be essential. I also worry that the BIOS might be like the EeePC with ACPI errors and PXE boot issues. I wont know until I actually test this hardware. MSI Wind PC Intel 945GC 1 x 200Pin Intel GMA 950 Black Barebone Specifications of this machine sound to be really good, especially at the listed price of only $149.99 + cost of RAM. You even get the SD slot, although only VGA. For ~60% the price of the two machines above it seems to fit all desired requirements except one. This machine was intended to insert a 3.5" hard drive, PCI card and even a 5.25" optical drive. The case is TOO BIG AND BULKY. This might be a very competitive product if it came in a sleeker case like the Asus Eee Box though. BIOS compatibility of ACPI and PXE booting must also be tested. Anyone know of other Intel Atom based thin clients or "nettop" tiny desktops on the market? Background Info (or Why Intel Atom with 2GB RAM?) A big problem in the past of all thin client or diskless workstation form-factor machines has been poor Linux driver support for the hardware, especially poor X drivers. All past thin clients have various chipsets like VIA UniChrome, AMD Geode or SIS that are just substandard. They have buggy drivers that became even worse as those companies failed to put forth resources and dedication to keep up with upstream development. While we now see both VIA and AMD trying to improve their relationship with upstream kernel and X communities, they have a significant amount of catch-up work ahead of them to match Intel's success from their past years of upstream development. Intel Atom appears to be a significant step ahead of all competition for Linux thin client and diskless workloads. The upcoming VIA Nano appears to have some promise with more performance and power efficiency. But again the quality and capability of the embedded video and graphics drivers will make or break its suitability for the Linux diskless workstation market. I would love it if VIA Nano becomes credible competition to Intel Atom. A recent development in LTSP is Local Applications support. This is a "hybrid" client model where diskless clients login to a server as thin client desktop session. But certain graphic intensive applications like Firefox or Google Earth run on the local machine displaying to the same X display. Users do not notice any difference as it is a seamless desktop experience. This has become increasingly important because media-intensive uses like Youtube require an unmanageable level of bandwidth (40-80mbit/sec per client!) when operating in a thin client model. Intel Atom's enables for the first time an affordable ultra-tiny machine with usable Linux 3D graphics. To support this new "hybrid" client model, even 2GB of RAM per client is now affordable at well under US$400 total per client. News of dual-core Atom coming soon makes this platform even more capable as a decent diskless workstation for mass deployments. (Comment on this)
Monday, August 11th, 2008
10:19 pm	How to make Flash 10 on Fedora not Crash Instantly Several new crashers for Fedora 10 happened in the last few weeks. nspulginwrapper-1.1.0 hit Fedora 9 updates with its new support for "windowless" WMODE. It turns out that the previous nspluginwrapper version did not support WMODE, so it was hiding an instant crasher (upstream bug) where firefox would die even with the protection of nspluginwrapper. Reportedly the next upstream release of firefox-3.0.2 should include this patch, but meanwhile lots of people are affected. Please test the builds of xulrunner from here. This patch against nspluginwrapper-1.1.0 is also necessary to use Flash 10 with WMODE without instant crashes. This was pushed as nspluginwrapper-1.1.0-5.fc9 in Fedora 9 updates a few days ago. Earlier today (August 11th 2008) Adobe released the release candidate of Flash 10, version 10.0.0.569. The big changes in this build are requiring external libraries NSS and libcurl.so.3. RHEL4 and RHEL5 have libcurl.so.3 so they are apparently fine, but Fedora 8+ have libcurl.so.4. It turns out that Debian somehow has been shipping with a symlink from libcurl.so.3 to libcurl.so.4 after they found that the ABI is actually the same. Updates to curl with a similar link (actually an empty shared library linked to libcurl.so.4) are heading to Fedora 8, Fedora 9 and Fedora 10. If you cannot wait, grab the latest build for your Fedora here. Unfortunately there seem to remain other crashes, but they are at least more difficult to reproduce. If you do figure out a consistent reproduce procedure for additional crashes please reply here. http://macromedia.mplug.org/ I just rewrote the old Flash RPM site into a permanent "Tips and Tricks" page where the latest Linux workarounds for these issues will be written down. (Comment on this)
12:31 am	Fedora 9 Live LTSP Server, Beta 1 Below is a LiveUSB or LiveDVD image containing Fedora 9 LTSP server and the client chroot pre-installed and pre-configured.  This is the easiest way to get started with Fedora LTSP5, or to try it without installing  onto your hard drive. Simply boot this Live image and follow the simple README, and you can within minutes demo serving of thin clients.  You can also install onto a hard drive to create a permanent server.  It is all surprisingly self-explanatory once you have booted to the Desktop. https://fedorahosted.org/k12linux/ LTSP on Fedora 9 w/ Updates is currently considered to be production ready, and development of improved features continues rapidly.  Check out our homepage for the latest news and updated instructions. Download http://alt.fedoraproject.org/pub/alt/ltsp/beta1/i686/ This image is based on Fedora 9 w/ Updates as of August 8th, 2008.  It seems to work great for me in limited tests. How to Use LiveUSB? https://fedorahosted.org/liveusb-creator From Windows you can use LiveUSB creator to make a USB stick bootable containing this Live image.  From Linux, within the ISO is the livecd-iso-to-disk script which you can use in Linux to make a bootable USB stick.  It is highly recommended that you use a persistent overlay file of at least 900MB because this reduces the amount of memory needed for your demo.  For this reason you should have at least 2GB free on your USB stick (~920MB image + 900MB overlay). LiveDVD works, but is not recommended unless you have at least 2GB RAM for your demo. Be warned that serving clients from a Live image will eat up your overlay and RAM rapidly, causing the demo to eventually fail.  It should be fine for a number of client logins.  You should install to your hard drive if you want to do more than just a quick demo. FAQ Why is this not called K12LTSP? It is the plan for Fedora LTSP5 technology to be the successor of Eric Harrison's highly successful K12LTSP distribution.  However we had planned on changing the name to "K12Linux" to be friendlier sounding and easier to pronounce when people explain it at educator conferences.  Unfortunately the naming issue remains a bit uncertain because we have not yet received ownership of k12linux domains from the current owner.  Warren's fault for getting busy and forgetting to follow up in past months. Is this the only way to install a Fedora LTSP5 server? This Live LTSP Server image is only a convenient way for new users to get started with Fedora LTSP5.  Note that it is always possible to enable LTSP5 on any existing Fedora 9 server by following the instructions on the above homepage. Why not LiveCD? LiveCD was not possible because we simply cannot fit Server, Client and apps onto a single disc.  If all you have is a CD drive then your hardware is unlikely powerful enough to serve as a LTSP server.  In any case you should be able to install from the LiveUSB without dealing with discs at all. Release Plan for Fedora 9 Live LTSP Server If we do not run into any serious problems, I plan on doing a Release Candidate ISO spin on Thursday, August 21st.  If no problems are found with the RC image, then it will be redubbed as "final" without any changes. I hope to get Fedora branding onto the login screen, and the naming issue straightened out before this release.  Delays in the final release may occur due to this branding stuff.  In the mean time, this Beta 1  spin seems to work well for me.  Please give it a try and let me know. https://www.redhat.com/mailman/listinfo/k12linux-devel-list Please send questions or comments to the k12linus-devel-list. (Comment on this)
Monday, August 4th, 2008
10:17 pm	Civilization Revolution 1.20: Still Very Broken I last reported on Civilization Revolution 1.10 two weeks ago. Now August 4th it is nearly a month since the American release, and two months since the European release. Three days ago they released the second patch, version 1.20, which was supposed to correct the absolutely broken  multiplayer matching system of 1.10. I can report with strong certainty that the game remains very broken on the Playstation 3 for online gaming. While the game is quite fun in single player, it is quickly too easy even at the highest computer difficulty. The 1.20 game sometimes rarely manages to connect to other players online, and then it can be very fun when it works.  A company rep on the 2KGames forums blames user's networking settings, routers and firewalls for these problems. However a multitude of complaints in several threads indicate a consistent story of dismal failure that agrees with my own frustrated attempts to play this game. Even after users port forward all of the suggested ports, put their PS3 in the DMZ (all ports forwarded) or even plugged directly into the cable modem, they are still unable to host or connect to most existing games.  Indeed my own experience is exactly this.  I estimate roughly 10% success rate in joining other people's games, and 5% success in hosting in dozens of attempts in the past few days since the 1.20 patch. All of these users (Including myself) reports having no problems playing online with every other PS3 multiplayer game. I successfully got into a multiplayer game maybe 9 times since the 1.20 patch.  Of those games, two games survived through completion.  7 of the 9 games failed either with complete deadlock of the PS3 or the game getting stuck with "rotating gears", seemingly due to some kind of network or server issue. Furthermore there seems to be no explanation from the company why the game does not take advantage of automatic UPNP NAT traversal for port forwarding.  It is a complete failure to expect end-users of console games to manually port forward in order to play online. ~2 months since release and after a fresh patch, it sure seems from this perspective that this company really screwed up.  This is a strong recommendation to AVOID this game.  It is currently far too broken and frustrating to be worth $60.  If they manage to fix these problems and use automatic UPNP NAT traversal the game might be worthwhile. I had so much hope for this game, yet I have never purchased a game this broken before.  I especially dislike the attitude expressed by the company blaming users instead of apologizing.  I have never traded in a game before, but I am contemplating doing so now. (Comment on this)
Tuesday, July 29th, 2008
7:02 pm	Full Report: LTSP Hackfest Portland 2008 The Linux Terminal Server Project (LTSP) has for ~9 years enabled Linux to act as a terminal server to diskless thin clients. LTSP has saved schools and businesses countless million, enabled or expanded access to technology to tens of thousands of schools globally, and enabled millions of otherwise useless obsolete computers to be recycled as thin clients. Today LTSP through various distributions like Debian, Ubuntu, Fedora and CentOS is very likely the largest by far exposure of Linux desktops to the general public. Long before OLPC, LTSP has and continues to be a major Linux Desktops for education success story. Now LTSP is becoming a shining star of cross-distribution coordination. Developers from Debian, Fedora, Gentoo, OpenSuSE, Ubuntu and others met in Portland, OR for the LTSP Hackfest. Of particular note was the efficacy of gobby to collaboratively write notes and conduct group review of code. It was great to see developers on 5 different distributions using a shared tool to work on common problems. We got more work done at this Hackfest than any previous event. Beyond just implementation of new code, we defined a clear plan for further implementation of features and bug fixes in the coming months. DisklessWorkstations.com has been the primary sponsor of LTSP upstream development for many years now. They covered a substantial amount of food for the event including the grand banquet on Day 1. Red Hat chipped some money to sponsor other meals. The coding stars of the event were Gideon Romm of Symbio Technologies , Scott Balneaves of LTSP upstream, and Ryan Niebur of Debian. Warren was mostly in a release engineering role, plus organizing the action item list, fixing a few other bugs, and reporting here what happened during the event. Also special thanks goes out to Freegeek of Portland, OR for use of their unique facility and for the tremendous service they do for the community in computer recycling and expansion of Linux desktop awareness in the community. Link to Day 1, Day 2, Day 3 notes in raw detail. LDM - The LTSP Display Manager LDM is a greeter and login screen comparable to GDM or KDM, except it uses the name and password to launch an SSH session to a target host. Upon successful authentication, it launches a Linux desktop session like GNOME or KDE. Scott Balneaves led the group in a live code review of the proposed LDM cleanup. The group suggested a few tiny improvements while others quietly added code cleanups in other parts of the same source file opened in the shared gobby session. It quickly was merged into ltsp-trunk and further testing ensued. It was important to get this merge done ASAP since it contains lots of code cleanups and simplification, that enables addressing other LDM action items. LDM Cleanup Troubles After Scott Balneaves' major code cleanup of ldm was committed to trunk on Day 1, the Debian and Ubuntu developers were having trouble getting logins to work. It seemingly worked on Warren's Fedora 9 laptop and Ryan52's Debian laptop but not on the slower laptops of Oliver and Vagrant. We eventually discovered that Scott introduced some timing dependent race condition. Scott battled that problem through most of Day 2 and eventually checked in a fix for Day 3 of the Hackfest. A separate cleanup fix submitted on ltsp-developer list eliminating the "kill -1" hack introduced a new issue causing logout to stall on XFCE and IceWM sessions, so it was backed out late Day 3. The only remaining regression then was local storage devices (like USB sticks) would no longer automatically unmount at logout. Vagrant fixed this at the very last moment of Day 3, we tested it quickly and tagged it as ldm-2.0.9. Gideon began implementing splitting of X launching from the ldm binary into the shell scripts that launch ldm, but that is not yet complete as of this writing. The LDM breakages introduced on Day 1 were very frustrating for the Debian and Ubuntu developers. But thanks to the group's effort we managed to leave the Hackfest with no known regressions and an improved foundation upon which to build further LDM improvements planned during the event. Local Apps Implementation We spent the final hours of Day 1 going through the "Local Apps" plan. Local apps enables running of certain applications on the thin client itself, which is tremendously beneficial with otherwise bandwidth intensive media applications. For example, Firefox playing Youtube can easily overwhelm a 100mbit network with only two or three clients. For this reason there is a strong desire to be able to offload video intensive applications like Firefox to the client, creating a "hybrid" client. The recent openssh-5.1 release included statvfs in the sftp-server. sshfs (somehow) already had statvfs and in simple tests it seems to work. Prior to arriving at the Hackfest, Gideon Romm had already thought out the entire local apps implementation. Warren suggested a simplification of the user/group portion to avoid pulling in yet another dependency of libnss_extrausers. A theoretical but untested implementation was written by Gideon by the end of Day 1. xrexecd: xatom signalled remote execution daemon xrexecd is a daemon that runs on the client immediately after LDM login. It connects and listens on the X server for certain xatom messages sent by the user session (from the terminal server) from the xrexec script. Before the 2nd day of the Hackfest, Scott Balneaves did a major cleanup of xrexecd in ltsp-client. Fedora built it in the ltsp-client packages before, but nobody used it yet since other infrastructure (sshfs homedirs) were not yet possible. xrexec example: A user might click on a special firefox launcher on the terminial server, which signals xrexecd to launch firefox on the diskless client. It is highly desirable to run firefox on the client itself instead of the terminal server in order to minimize the bandwidth of media-heavy use like Youtube. Google Earth may be another popular application to run on the client instead of terminal server. xrexecd exits immediately on Fedora because it is unable to connect to the X server due to some Xauthority permission problems. Scott and Warren began a little debugging and found some some differences in Fedora with stricter permissions. More work is required on xrexecd to properly handle XAUTHORITY access in both X forwarded and non-forwarded modes of operation. Warren hopes to get this working in the next week or two. It is very exciting for us LTSP hackers to have this soon working since this has been the most desired feature for years. sshfs mounted $HOME seems to work Gideon Romm's rc.d/S01-localapps script merged early Day 2 seems now to be stable. It handles mounting of the user's home directory with sshfs via the already existing ssh tunnel from the LDM session. It also replicates the minimum of passwd and group entires to allow local apps to use that home directory. Warren confirmed that the new openssh-5.1 on the terminal server successfully allows sshfs-mounted home directories to work with firefox. It is a little slow, but it seems to work, although a few crashes occurred that we still need to check out. Fedora was the first to run firefox as a LTSP local app, but it was run manually due to the aforementioned Xauthority permission difficulties in xrexecd. Fedora can soon benefit from sshfs homedirs with openssh-5.1, where it is currently awaiting Bodhi karma votes in Fedora 9 updates-testing. Live LTSP Server almost there Fedora Live LTSP Server is a LiveUSB or LiveDVD where you can host network boot thin clients or diskless workstations without installing anything to your hard drive. Since it has a client chroot in addition to the server OS + desktop applications, it is impossible to fit within a LiveCD. This image is mostly useful for quick demos, testing purposes, or to install a fully configured and ready-to-go Fedora LTSP server onto a hard drive. This image contains pretty much just Fedora Desktop LiveCD, but with ltsp-server and the client chroot pre-installed in /opt/ltsp/i386, with all the configurations and services to serve LTSP server on the ltspbr0 bridge pre-enabled. Warren got very far toward making this work during the Hackfest. Two other Hackfest attendees went home with a 2GB USB stick containing the first attempt, which worked after a few manual tweaks saved to the LiveUSB persistent overlay. Further work ensues to make it release-ready, which includes: 1) Adding OpenOffice and a few education type applications 2) Figuring out a clear way to explain to the user how to configure the networking, since Ethernet bridging is not yet supported by NetworkManager, it is not exactly straight forward like a normal LiveCD. LTSP Documentation Writers Team Founded Brian Tilma founded the ltsp-docwriters team. Writing documentation is a great way to get involved with LTSP if you are not a coder. If you are interested in helping please talk with us on the ltsp-developer list or #ltsp on irc.freenode.net. Gentoo Almost Ready Donnie Berkholz and Johnny Robeson of Gentoo attended this Hackfest. They made a lot of progress toward making Gentoo working with LTSP5. They struggled primarily with their genkernel initramfs tool. Donnie informed the ltsp-developer list a day after the Hackfest that it is now working. It seems we will soon see LTSP5 easily working in Gentoo. Ubuntu Intrepid uploaded new upstream code Oliver Grawert (ogra) spent most of the Hackfest getting reaquainted to the many changes that happened upstream since he forked off for the previous Ubuntu stable release. By the end of Day 3 he had new upstrea code uploaded toward Ubuntu Intrepid. He probably now needs testers to give him feedback. Next Two LTSP Hackfests There is already planning underway for the next two LTSP Hackfests. The next will be in Maine tenatively scheduled to be around October 30th. This makes it roughly a year after the previous Hackfest in Southwest Harbor, Maine. If you have any interest in LTSP or diskless clients, I can highly recommend hanging out at this next Hackfest. Accommodations are very affordable in the cozy Seawall Motel during the tourist off-season near Acadia National Park. We plan on eating the local food, which includes lobster and other delicious seafood. (Food is very important to LTSP events.) Jim McQuillan, founder of the LTSP project has made it a goal to send all of the LTSP developers to the FISL conference in Brazil on June 24th, 2009. It will be the 10th anniversay of both FISL and the LTSP project, and LTSP is very big in Brazil because of the tremendous enabler it has been for education. This should prove to be the biggest LTSP event ever. We have no idea how we will pay to send LTSP developers from the various distributions to the event, but we have 11 months to figure that out. While Warren has the good fortune of having Red Hat pay for his occasional visits to Hackfests, most of the other developers pay out of their own pockets to work on the public good. Perhaps we need to organize some kind of fund raiser. If you have any questions or comments for the LTSP project please reply here. We would be happy to answer. (Comment on this)
Monday, July 21st, 2008
9:14 pm	Civilization Revolution is Very Broken UPDATE: I wrote new commentary on August 4th, 2008 about the 1.20 patch. Summary: The game remains very broken, and the company blames users instead of apologizing. July 8th Civilization Revolution was released in America. I pre-ordered it from Amazon because the demo on PS3 was fun. I didn't receive it until July 14th. I can report that while the game can be quite fun, the game suffers from numerous crippling bugs and design problems. For this reason I must currently recommend AGAINST purchase at a steep price of $60. It pains me to write this negative recommendation because I am a huge fan of the Civ series and hope to have fun playing with my friends. Perhaps I will change my mind later if they fix most of these problems. HUGE DESIGN PROBLEM: NAT TRAVERSAL - CRIPPLED BY DEFAULT NETWORK GAMING A significant issue evident even during the PS3 demo was a problem in the game's network handling in Multiplayer. This is where you do a search for online games, and it tells you it failed to host, check your network connection. It turns out that you need to manually forward ports to your PS3 in order for the game to work reliably multiplayer (it seems it is possible to play without forwarding if you connect to another hosting player who is forwarded.) This demonstrates an immense failure of design. So many other games work just fine without manual port forwarding. Other multiplayer PS3 games like Warhawk work automatically with UPnP NAT traversal for the PS3 to report "NAT Type 2" in its network test. Why can the PS3 network test and other games like Warhawk work reliably with automatic UPnP NAT traversal and Civilization Revolution requires manual port forwarding? While my engineer geekiness can handle manual port forwarding, this is a terrible expectation to make of ordinary users. It is a great way to limit the attractiveness of your game when it fails by default. There appears to be some unconfirmed speculation that instability in the game hosting is due to the company decision to use the GameSpy network. From what I can tell GameSpy is only a matching service, while players require manually forwarded ports in order to play against each other. The XBox 360 version of Civ Rev apparently does not have this problem because the games are hosted on central servers. CRIPPLED MULTIPLAYER SEARCH AT AMERICAN PS3 RELEASE Immediately upon release they issued a patch to 1.10 which horribly crippled Multiplayer, a large reason why this game has any appeal. The 2K Games Forums are full of people screaming about this and numerous other bugs, and reportedly they are working on the 1.20 version to fix searches. What seems to be the issue is 1.10 incorrectly made all game searches look for "Scenario" games instead of ordinary games. This means that everyone attempting to search will find nobody, leading players to believe there really are no other players. Users can find other players by downgrading to 1.00 (delete the game, run it and refuse upgrade). However if they start a game with another user running 1.10 it is impossible to play due to resynching errors. It is only possible to play a multiplayer game currently if all players in the game are running the same version. A representative of the company said 1.20 will correct the search so players can easily find each other, and also prevent mixed version games. This bug alone is an incredible embarassment that will be very costly to the company. RACE CONDITIONS AND DEADLOCKS The forums are full of complaints of deadlocks and other quirks that can happen during a multiplayer game. I personally experienced the "dancing units" bug where combat can cause the game to get stuck. It seems to happen only rarely because an asynchronous event of another player must occur at an instant where you enter combat. In my case another player settled a new city in an adjacent spot to where I was attacking another player. Combat ensued, ended, and the game froze. Judging by the forum posts, it seems the company has not admitted this to be a confirmed problem, and it is not scheduled to be fixed in vesion 1.20. MISGUIDED PRIORITIES AT COMPANY Upon American release the company seemed very excited about talking about their upcoming Downloadable Content that players can purchase. This is extremely annoying to players on the 2K Games Forums and myself that they would prioritize the creation of downloadable content to squeeze more money out of players when the underlying game is so demonstrably broken. POOR COMMUNICATION/NEWS IN GAME Upon running the game and connecting to the network, there is no indication of news bulletins from the company to players like exists in other online console games like Warhawk. It might be helpful to have something that can display tips to users to warn them of the crippling NAT traversal problem and what to do about it. Or to talk about new things fixed in version upgrades. I soon leave on a business trip and return next week. I hope they issue the next update soon. I will reassess if the game is improved, and update this blog entry. (Comment on this)
Sunday, July 6th, 2008
1:42 am	Radicals Dreamer on Guitar Last weekend I took my first guitar lesson and bought a classical guitar Ibanez G850. This weekend was my second lesson. I am finding it to be very challenging and sometimes frustrating when I'm not progressing fast enough. But I feel inspired by various solo guitar performances on Youtube. Today I pre-paid and committed myself to 12 more lessons for a discount, and to give myself further incentive to keep practicing. My new favorite guitar video. Justin Lincoln did a superb job in arranging the rhythm guitar and singing parts of the Chrono Cross ending theme into a single guitar solo complete with video and tabs. I hope that one day I will be able to play this song. (Comment on this)
Sunday, June 22nd, 2008
1:02 am	素敵だね (Isn't It Wonderful?) I did some random web surfing to get my mind off the pain. I really like the musical themes from the Final Fantasy series.  I was delighted to discover that people have been putting videos of themselves playing guitar arrangements on Youtube.  Youtube a ton of this stuff.  Amazing platform of creativity it is.  A few that I liked... Acoustic Guitar: Tifa's Theme (FF7) Main Theme (FF7) Breezy (FF8) Eyes on Me (FF8) 600AD (Chronotrigger ... not FF but made by the same company and with equally memorable themes.) Melodies of Life (FF9) Check out this guy's playlist.  He arranged and played 32 FF songs for solo guitar.  Some are better than others, but I appreciate this guy's work. Electric Guitar: FF7 Battle Theme Chocobo Theme (This guy is playing bass and lead AT THE SAME TIME.  His attitude is funny too.  He has a few other videos where he does Zelda and Tetris with two guitars as well.) Chronotrigger Battle Theme (Entertaining watching the small kid play huge instruments.  First he lays down the bass line in a loop, then goes nuts on the guitar part.) An attractive Japanese girl put a lot of effort into dressing up as the character Yuna from Final Fantasy X and singing 素敵だね (Suteki dane), the theme song from the game.  I do admit that her face distracted me from a few anemic notes (...), but she did a good job reproducing the singing of Japanese folk singer Rikki.  As a former musician I have a lot of respect for the skill and especially *guts* it takes to post yourself on Youtube. During my most recent visit to Tokyo I was delighted to find that game music like this was somewhat mainstream, in the rotations of some radio stations and cafes.  I heard the original version of this song and a few instrumental arrangements of other Final Fantasy themes. 素敵だね acoustic guitar version.  This guy does a good job playing this commercial arrangement from a song book. This makes me seriously consider attempting to learn acoustic guitar.  How do I begin?  (I've done only woodwinds in the past.) (Comment on this)
Thursday, June 19th, 2008
4:02 pm	Keyboard Plan FUDCon Boston 2008 Hackfest The keyboard configuration in Linux has long been a huge mess of too many different ways of doing the same thing in different places and inconsistencies.  (Please correct me if I have any details wrong.) * Today there are two sets of keyboard layout data, kbd for the console and xkeyboard-config for X. * /etc/sysconfig/keyboard configures only the console keyboard layout. * When X runs, there is no connection of the console layout and X keyboard layout. * xorg.conf can specify a layout and options to use from xkeyboard-config (an upstream package with tables).  (Fedora/RHEL doesn't use this AFAIK.) * Or gdm reads a user-configured layout from ~/.dmrc, which can't always work because the home directory is not available in some cases before login. * After you have logged into GNOME, gnome-settings-daemon reads an optionally configured keyboard from gconf and uses it if available. * (Do other desktops like KDE have something else entirely?) With developers together during FUDCon Boston 2008 Hackfest, Jeremy Katz, Bill Nottingham and Adam Jackson had a chance to brainstorm a plan to improve this confusing mess. * We need to throw out the entirely separate kbd table.  Console keymap setting needs to happen from the shared xkeyboard-config.  ckbcomp from Debian exists.  Jeremy Katz is looking into adapting or rewriting it in C because there is some discomfort with the current perl implementation. * ajax already has some code to have X  set  the keyboard layout itself.  He plans on making it happen from whatever HAL tells it, which comes from /etc/sysconfig/keyboard on Fedora This still leaves the inconsistency of two separate user-specific keyboard layout options in ~/.dmrc and gconf.  But these at least do not (at least should not) happen if a user has not configured it. (Comment on this)